So the model here seems to be, people coming out of near-death experiences have these memories, and while they're likely not "real", they're a record of some sequence of cognitive states, and the puzzle is, how can we detect these cognitive states? There seems to be an underlying assumption that the memories are a faithful chronological record of something, and the investigation is, what is the something -- what is the brain recording while it's apparently inert.
This may well be right, they seem to have good evidence of apparently-inert brains being not-so-inert, so at this point I suppose I'm quibbling.
But the part I have never understood about discussions of near-death experiences (IANAneurologist) is, why do so many of these stories assume that the memories people wake up with were created during the apparently-inert time? It's true that the memories are subjectively of long duration, people report that they remember spending a lot of time flying towards the light or conversing with the angels, but surely they can be sincere without being right.
We know a fair amount now about how memories can be manipulated, and how recollections depend on the environment -- memories are very slippery things. So, isn't it possible that, during the apparently-inert period of a near-death experience, the brain actually is inert, and not forming memories, and that at the time of recovery, during which there is plenty of obvious brain activity, the memories are all formed in a brief period, but with the subjective sense of having taken place over a longer period? This means the memories are basically wrong, but this seems to me to be a much lower bar to clear than requiring chronologically faithful memory construction in quiescent gray matter.
Well, in fairness, it was prepared for the Queen, not necessarily by the Queen. It was her advisers who imagined the solemn and awful duty falling to her.
As you hint at, it's the libdvdcss capability that's the main problem under anti-circumvention provisions of the US DMCA.
You can get versions of VLC which only use FOSS and patent-unencumbered codecs. Debian used to (maybe still does, I haven't looked in a while) make this distinction pretty clear, the "main" packaged VLC was unencumbered, and you had to go outside the main package tree to get the other stuff.
So, in most practical installations, you're right, but it's not literally true that "VLC is illegal in the US."
So the article and summary hint at a common problem -- "the ministry has its own system for... sharing documents", which "doesn't always function well outside of Japan". I've seen this in more than one enterprise, where the IT guys meet the need of users to securely move data around by buying or building a secure solution, and they pay very careful attention to the security, but less attention to the usability. Users will go for ease-of-use every time, and aren't thinking about security, so mistakes like this happen.
The obvious solution is to make the secure system easy to use, but usability itself is hard to get right, secure usability is very hard.
As a naturalized US citizen who actually took a small quiz on this, I am honor-bound to point out that the fine quotation you have provided is actually from the Declaration of Independence, and not the Constitution. While it certainly reflects the aspirations of the founders, and may well represent my or your best hopes, it's not actually the law of the land. The constitution is clearer about its jurisdiction.
Any host-based intrusion detection system will have a hash of the executable, and will report when it changes. This is not some new cutting-edge security precaution, it's routine for many, many installations.
Seriously, he's a lawyer, in what particular does he think the rejection is wrong?
The nearest thing to a substantive accusation is that the examiner is simply rejecting the application because he's lazy and that's easy. But it's my understanding that, in fact, patent examiners face a lot of pressure to approve applications, which is faster and easier than rejection, because it takes less effort to justify approval, and because approvals don't generally get appealed by the applicant. So while I am sure laziness afflicts patent examiners from time to time, it's not obvious that this is an example.
As for "doing his job", his job is not to approve applications, it's to examine them and make a determination. Rejection is one possible outcome, and is not by itself proof that the job wasn't done.
So, yeah, faceless bureaucrats are lazy and stupid, ha ha. Tell me again what problem you solved by making this assertion?
I've been using it for a long time, too, it's a perfectly respectable choice, and if I had to use it for ten more years, that would be OK.
However, particularly for back-up systems, I am ready for snapshots and block-level deduplication. I tried to deploy something like this with XFS over LVM a few years ago, but discovered that the write performance of LVM snapshots degrades rapidly when there are a lot of them, and it helps a lot if you can guess the size in advance, which is hard. There's also a hard limit of 255 snapshots, but in our environment, performance became unacceptable before we got anywhere near that.
You're right that XFS "ain't broke", but I for one am ready for more features.
What you say is likely true for almost all users, but for server management, the network transparency features that come with server-client separation are a huge asset. My own "use-case" is that I frequently need to install commercial scientific software on remote headless systems, e.g. the head node of a computational cluster in the server room. These installers invariably have GUIs, which I use by SSH-ing into the box with a forwarded X connection and just running it.
There are other ways to do this, of course, you can use some kind of remote desktop scheme to accomplish the same goal, but you don't actually need the whole desktop, you really only need to operate the remote GUI on your existing local desktop. X can do this, Wayland (and Windows and Quartz) sacrifice this in order to have better local display performance.
I also worry that it's part of a general trend towards more monolithic software, and towards doing less in order to do it better. Unix (and Linux) were initially attractive to me because of their mind-set of having a good set of powerful, conceptually simple tools that I could chain together to accomplish my goals. Now, it seems like I'm seeing more and more conceptually complex, monolithic applications that are very, very good at solving the most frequent use case, but are somewhere between useless and harmful if you try something the developer didn't anticipate, because it's a niche requirement or a corner case. I'm starting to miss systems that worked in the corner cases.
Robert Zubrin, the "case for Mars" guy who seems to have thought a lot about months-long space journeys, believes that low-gravity bone loss can be mitigated by exercise. His data point is Shannon Lucid, who spent 179 days on the Mir space station, rigorously followed the prescribed exercise regime, and came back in significantly better physical condition than other members of her crew, who weren't as disciplined with their exercise regimes.
Even if he's wrong, this is a problem to be solved, rather than a reason not to try.
I suppose the confusion arises because of G. Washington's investments in trying to drain the Great Dismal Swamp, but this on the Virginia-NC border, not the site of Washington DC.
> No one fucking cares, I know this because... its going away and no one is saying loudly 'we can import your google reader feeds, move to us!!!!!'.
Actually, that's exactly what feedly was saying on their home page last night. They seem to have a good feature set, and run on all the platforms I care about, but their servers buckled under the load yesterday, so maybe not.
As far as I can tell from the article, the basis of the complaint is that vendors object to the fact that searching on their brand name or model name brings up stuff that's not theirs, and they believe that having these search results show up confuses consumers about who made the products in the search result.
So, if this is the case, then it's like, I go into a physical store, and say to the salesman, "I would like to buy an Apple laptop computer," and the salesman produces a computer, and says "Here is an Inspiron laptop computer, it has many wonderful features." The salesman neglects to mention that the Inspiron is an alternative to, rather than an example of, an Apple computer. The accusation is that the salesman is trading on Apple's good name to sell non-Apple merchandise.
It's similar to when people complain about sponsored search results not being easily distinguished from non-sponsored results.
One is to disallow password authentication via SSH. Then you can have weak passwords locally on the machine, and use public key authentication for remote access.
A second one is to only allow remote access to a special account with a long password, and then, when logging in remotely, su to the main account with the short password. This is a bit brittle, but would work.
A third is to re-examine how you're using your system -- you probably don't actually need to supply passwords all the time. There are other distros besides Ubuntu, and, contrary to what you might have heard, logging in as root to do system maintenance is both reasonable and allowed.
Having negative temperatures be "higher" than positive ones actually makes a lot of thermodynamic sense. For one thing, it lets you preserve the notion that heat naturally flows from hotter things to colder things.
Formally speaking, it's more natural to think in terms of the inverse of temperature, 1/T, sometimes called beta. In the limit of very large positive beta, that's nearly absolute zero, and is the low-energy end of the spectrum. A beta of zero is full disorder. Negative beta corresponds to high energy systems that nevertheless have some order, so that the concept of temperature can be (formally) defined.
It confounds your (well, my) intuition, because "ordinary" systems generally obey the rule that the more energy they have, the greater the number of states the system can be in, but that's not an actual law of physics, it's just the usual case.
It's actually not quantum mechanical, at least not explicitly, but it requires that the system in question have an ordered state that is at a high-energy bound.
A classical system that does this is an array of magnetic dipoles in a magnetic field. When all the dipoles are aligned against the field, the system is fully ordered, and is in its highest-energy state. If you look at deviations from this state, what you find is that all of them increase the entropy (because the state is fully ordered), and decrease the energy (because the state is an upper bound for energy), which means that the ratio of delta-E over delta-S, or dE/dS, is negative. Thermodynamically speaking (again, classically), dE/dS is (one definition of) the temperature T.
It's hard to tell if this is secretly really a quantum effect -- my example uses magnetic dipoles, which are arguably a quantum thing, either because of spin or because electronic orbitals that don't collapse because they're quantum states. But you don't need quantum mechanics in principle -- what you need is a low-entropy, high-energy state of the system with nearby states that have more entropy and less energy.
I can't think of a fully classical example that does this, but someone clever might be able to.
I actually got this negative-temperature question on my Ph.D. qualifying oral exam, and I totally kicked its ass.
So there is a trope in the engineering world that the safest reactors are the ones that are confined to paper studies, or, to put it more timely, to PowerPoint slides.
It's true that the LFTR reactors don't have the same failure modes as the pressurized light-water reactors, but they still have the same basic issue, namely that there is a very large amount of power-generating capacity in a relatively small volume. Even pebble-bed reactors, similarly touted as "intrinsically safe" during their design phase, have had a radiation-release accident -- scroll down to "Criticisms of the design" on that Wikipedia page. The lesson (which I learned from Charles Perrow and Fukushima) is that complex systems with high power densities are intrinsically hazardous, because unexpected interactions (which arise from the complexity) tend to be highly destructive (because of the power density). LFTRs are less complex, and so less dangerous, than PLWRs, and that's good, but it doesn't make them safe.
The stupid cliche you hear over and over again is true -- safety is a process. You can design reactors so that the safety process is easier to implement, but what actually makes things safe is conservative management schemes that retain the redundancy and margin for error that the process demands, and not cutting them out because of the money, or, worse, because of complacency induced by faith in the design.
There's another industrial safety joke, particularly applicable to complex systems -- accident analysis consists of filling in X and Y in the phrase, "Nobody imagined X could happen whlie Y was true."
My experience with my Nexus 7 is that getting an Android tablet from Google means you get updates regularly. My Nexus 7 shipped with 4.0.4, and is now up to 4.2.1, in under a year. I expect the Google phones are similar.
Slashdot Mirror
So the model here seems to be, people coming out of near-death experiences have these memories, and while they're likely not "real", they're a record of some sequence of cognitive states, and the puzzle is, how can we detect these cognitive states? There seems to be an underlying assumption that the memories are a faithful chronological record of something, and the investigation is, what is the something -- what is the brain recording while it's apparently inert.
This may well be right, they seem to have good evidence of apparently-inert brains being not-so-inert, so at this point I suppose I'm quibbling.
But the part I have never understood about discussions of near-death experiences (IANAneurologist) is, why do so many of these stories assume that the memories people wake up with were created during the apparently-inert time? It's true that the memories are subjectively of long duration, people report that they remember spending a lot of time flying towards the light or conversing with the angels, but surely they can be sincere without being right.
We know a fair amount now about how memories can be manipulated, and how recollections depend on the environment -- memories are very slippery things. So, isn't it possible that, during the apparently-inert period of a near-death experience, the brain actually is inert, and not forming memories, and that at the time of recovery, during which there is plenty of obvious brain activity, the memories are all formed in a brief period, but with the subjective sense of having taken place over a longer period? This means the memories are basically wrong, but this seems to me to be a much lower bar to clear than requiring chronologically faithful memory construction in quiescent gray matter.
Any neurologists in the crowd care to comment?
They're completely unhackable!.
Soon they'll be mandatory in Enterprise deployments.
Well, in fairness, it was prepared for the Queen, not necessarily by the Queen. It was her advisers who imagined the solemn and awful duty falling to her.
Not only Yahoo's index, they're blocking indexing for Google and Bing also. Presumably via robots.txt or similar.
As you hint at, it's the libdvdcss capability that's the main problem under anti-circumvention provisions of the US DMCA.
You can get versions of VLC which only use FOSS and patent-unencumbered codecs. Debian used to (maybe still does, I haven't looked in a while) make this distinction pretty clear, the "main" packaged VLC was unencumbered, and you had to go outside the main package tree to get the other stuff.
So, in most practical installations, you're right, but it's not literally true that "VLC is illegal in the US."
So the article and summary hint at a common problem -- "the ministry has its own system for ... sharing documents", which "doesn't always function well outside of Japan". I've seen this in more than one enterprise, where the IT guys meet the need of users to securely move data around by buying or building a secure solution, and they pay very careful attention to the security, but less attention to the usability. Users will go for ease-of-use every time, and aren't thinking about security, so mistakes like this happen.
The obvious solution is to make the secure system easy to use, but usability itself is hard to get right, secure usability is very hard.
As a naturalized US citizen who actually took a small quiz on this, I am honor-bound to point out that the fine quotation you have provided is actually from the Declaration of Independence, and not the Constitution. While it certainly reflects the aspirations of the founders, and may well represent my or your best hopes, it's not actually the law of the land. The constitution is clearer about its jurisdiction.
Any host-based intrusion detection system will have a hash of the executable, and will report when it changes. This is not some new cutting-edge security precaution, it's routine for many, many installations.
Seriously, he's a lawyer, in what particular does he think the rejection is wrong?
The nearest thing to a substantive accusation is that the examiner is simply rejecting the application because he's lazy and that's easy. But it's my understanding that, in fact, patent examiners face a lot of pressure to approve applications, which is faster and easier than rejection, because it takes less effort to justify approval, and because approvals don't generally get appealed by the applicant. So while I am sure laziness afflicts patent examiners from time to time, it's not obvious that this is an example.
As for "doing his job", his job is not to approve applications, it's to examine them and make a determination. Rejection is one possible outcome, and is not by itself proof that the job wasn't done.
So, yeah, faceless bureaucrats are lazy and stupid, ha ha. Tell me again what problem you solved by making this assertion?
I've been using it for a long time, too, it's a perfectly respectable choice, and if I had to use it for ten more years, that would be OK.
However, particularly for back-up systems, I am ready for snapshots and block-level deduplication. I tried to deploy something like this with XFS over LVM a few years ago, but discovered that the write performance of LVM snapshots degrades rapidly when there are a lot of them, and it helps a lot if you can guess the size in advance, which is hard. There's also a hard limit of 255 snapshots, but in our environment, performance became unacceptable before we got anywhere near that.
You're right that XFS "ain't broke", but I for one am ready for more features.
What you say is likely true for almost all users, but for server management, the network transparency features that come with server-client separation are a huge asset. My own "use-case" is that I frequently need to install commercial scientific software on remote headless systems, e.g. the head node of a computational cluster in the server room. These installers invariably have GUIs, which I use by SSH-ing into the box with a forwarded X connection and just running it.
There are other ways to do this, of course, you can use some kind of remote desktop scheme to accomplish the same goal, but you don't actually need the whole desktop, you really only need to operate the remote GUI on your existing local desktop. X can do this, Wayland (and Windows and Quartz) sacrifice this in order to have better local display performance.
I also worry that it's part of a general trend towards more monolithic software, and towards doing less in order to do it better. Unix (and Linux) were initially attractive to me because of their mind-set of having a good set of powerful, conceptually simple tools that I could chain together to accomplish my goals. Now, it seems like I'm seeing more and more conceptually complex, monolithic applications that are very, very good at solving the most frequent use case, but are somewhere between useless and harmful if you try something the developer didn't anticipate, because it's a niche requirement or a corner case. I'm starting to miss systems that worked in the corner cases.
FTFS: "Three serial interfaces are available via the expansion headers." So it's a connector and a few minutes of soldering.
Robert Zubrin, the "case for Mars" guy who seems to have thought a lot about months-long space journeys, believes that low-gravity bone loss can be mitigated by exercise. His data point is Shannon Lucid, who spent 179 days on the Mir space station, rigorously followed the prescribed exercise regime, and came back in significantly better physical condition than other members of her crew, who weren't as disciplined with their exercise regimes.
Even if he's wrong, this is a problem to be solved, rather than a reason not to try.
Seriously? Emergency medical technician, aka paramedic. The guy in the ambulance who does the cardio-pulmonary resuscitation.
Obligatory link.
It turns out this hypothetical scenario actually was too extreme, it was set much too far in the future...
Furrfu!
> Government builds capital on swamp ...
It's not, actually.
I suppose the confusion arises because of G. Washington's investments in trying to drain the Great Dismal Swamp, but this on the Virginia-NC border, not the site of Washington DC.
> No one fucking cares, I know this because ... its going away and no one is saying loudly 'we can import your google reader feeds, move to us!!!!!'.
Actually, that's exactly what feedly was saying on their home page last night. They seem to have a good feature set, and run on all the platforms I care about, but their servers buckled under the load yesterday, so maybe not.
As far as I can tell from the article, the basis of the complaint is that vendors object to the fact that searching on their brand name or model name brings up stuff that's not theirs, and they believe that having these search results show up confuses consumers about who made the products in the search result.
So, if this is the case, then it's like, I go into a physical store, and say to the salesman, "I would like to buy an Apple laptop computer," and the salesman produces a computer, and says "Here is an Inspiron laptop computer, it has many wonderful features." The salesman neglects to mention that the Inspiron is an alternative to, rather than an example of, an Apple computer. The accusation is that the salesman is trading on Apple's good name to sell non-Apple merchandise.
It's similar to when people complain about sponsored search results not being easily distinguished from non-sponsored results.
There are several solutions to your problem.
One is to disallow password authentication via SSH. Then you can have weak passwords locally on the machine, and use public key authentication for remote access.
A second one is to only allow remote access to a special account with a long password, and then, when logging in remotely, su to the main account with the short password. This is a bit brittle, but would work.
A third is to re-examine how you're using your system -- you probably don't actually need to supply passwords all the time. There are other distros besides Ubuntu, and, contrary to what you might have heard, logging in as root to do system maintenance is both reasonable and allowed.
Having negative temperatures be "higher" than positive ones actually makes a lot of thermodynamic sense. For one thing, it lets you preserve the notion that heat naturally flows from hotter things to colder things.
Formally speaking, it's more natural to think in terms of the inverse of temperature, 1/T, sometimes called beta. In the limit of very large positive beta, that's nearly absolute zero, and is the low-energy end of the spectrum. A beta of zero is full disorder. Negative beta corresponds to high energy systems that nevertheless have some order, so that the concept of temperature can be (formally) defined.
It confounds your (well, my) intuition, because "ordinary" systems generally obey the rule that the more energy they have, the greater the number of states the system can be in, but that's not an actual law of physics, it's just the usual case.
It's actually not quantum mechanical, at least not explicitly, but it requires that the system in question have an ordered state that is at a high-energy bound.
A classical system that does this is an array of magnetic dipoles in a magnetic field. When all the dipoles are aligned against the field, the system is fully ordered, and is in its highest-energy state. If you look at deviations from this state, what you find is that all of them increase the entropy (because the state is fully ordered), and decrease the energy (because the state is an upper bound for energy), which means that the ratio of delta-E over delta-S, or dE/dS, is negative. Thermodynamically speaking (again, classically), dE/dS is (one definition of) the temperature T.
It's hard to tell if this is secretly really a quantum effect -- my example uses magnetic dipoles, which are arguably a quantum thing, either because of spin or because electronic orbitals that don't collapse because they're quantum states. But you don't need quantum mechanics in principle -- what you need is a low-entropy, high-energy state of the system with nearby states that have more entropy and less energy.
I can't think of a fully classical example that does this, but someone clever might be able to.
I actually got this negative-temperature question on my Ph.D. qualifying oral exam, and I totally kicked its ass.
So there is a trope in the engineering world that the safest reactors are the ones that are confined to paper studies, or, to put it more timely, to PowerPoint slides.
It's true that the LFTR reactors don't have the same failure modes as the pressurized light-water reactors, but they still have the same basic issue, namely that there is a very large amount of power-generating capacity in a relatively small volume. Even pebble-bed reactors, similarly touted as "intrinsically safe" during their design phase, have had a radiation-release accident -- scroll down to "Criticisms of the design" on that Wikipedia page. The lesson (which I learned from Charles Perrow and Fukushima) is that complex systems with high power densities are intrinsically hazardous, because unexpected interactions (which arise from the complexity) tend to be highly destructive (because of the power density). LFTRs are less complex, and so less dangerous, than PLWRs, and that's good, but it doesn't make them safe.
The stupid cliche you hear over and over again is true -- safety is a process. You can design reactors so that the safety process is easier to implement, but what actually makes things safe is conservative management schemes that retain the redundancy and margin for error that the process demands, and not cutting them out because of the money, or, worse, because of complacency induced by faith in the design.
There's another industrial safety joke, particularly applicable to complex systems -- accident analysis consists of filling in X and Y in the phrase, "Nobody imagined X could happen whlie Y was true."
My experience with my Nexus 7 is that getting an Android tablet from Google means you get updates regularly. My Nexus 7 shipped with 4.0.4, and is now up to 4.2.1, in under a year. I expect the Google phones are similar.
Same way I read everything, Pinky.